guideline
Guidelines for the diagnosis and management of hereditary
spherocytosis
P. H. B. Bolton-Maggs1, R. F. Stevens2, N. J. Dodd3, G. Lamont4, P. Tittensor5and M.-J. King6 on behalf of the General
Haematology Task Force of the British Committee for Standards in Haematology
1
Manchester Royal Infirmary, Manchester, 2Royal Manchester Children’s Hospital, Manchester, 3The Ipswich Hospital NHS Trust,
Ipswich, 4Royal Liverpool Children’s Hospital, Alder Hey, Liverpool, 5Patient Representative, and 6International Blood Group Reference
Laboratory, Bristol, UK
Summary
Hereditary spherocytosis (HS) is a heterogeneous group of
disorders with regard to clinical severity, protein defects and
mode of inheritance. It is relatively common in Caucasian
populations; most affected individuals have mild or only
moderate haemolysis. There is usually a family history, and a
typical clinical and laboratory picture so that the diagnosis is
often easily made without additional laboratory tests. Atypical
cases may require measurement of erythrocyte membrane
proteins to clarify the nature of the membrane disorder and in
the absence of a family history, occasionally molecular genetic
analysis will help to determine whether inheritance is recessive or
non-dominant. It is particularly important to rule out stomatocytosis where splenectomy is contraindicated because of the
thrombotic risk. Mild HS can be managed without folate
supplements and does not require splenectomy. Moderately and
severely affected individuals are likely to benefit from splenectomy, which should be performed after the age of 6 years and
with appropriate counselling about the infection risk. In all cases
careful dialogue between doctor, patient and the family is
essential. Laparoscopic surgery, when performed by experienced
surgeons, can result in a shorter hospital stay and less pain.
Keywords: spherocytosis, hereditary, splenectomy, child,
erythrocyte membrane.
Hereditary spherocytosis (HS) was described in 1871 and the
first recorded splenectomy was performed soon after. It is the
commonest cause of inherited chronic haemolysis in Northern
Europe and North America with a quoted incidence of 1 in
5000 births (Morton et al, 1962). However, studies of osmotic
Correspondence: BCSH Secretary, British Society for Haematology,
100 White Lion Street, London, N1 9PF, UK.
E-mail: jules@b-s-h.org.uk
RF Stevens sadly died prior to the final revisions of these guidelines.
fragility in blood donors suggest the existence of extremely
mild or subclinical forms (Godal & Heisto, 1981; Eber et al,
1992), raising the prevalence of HS to 1 in 2000 . HS has also
been found in other ethnic groups (in Africa, Algeria, Tunisia,
Egypt, Japan, North India and Brazil). There are only rare cases
reported in the black population. In the last 15 years significant progress has been made in the understanding of the
biochemical and molecular genetic basis of HS.
This guideline addresses several practical aspects:
1 Diagnosis
a. Which individuals should be screened for HS?
b. Which are the best diagnostic tests?
c. When is it necessary to measure membrane proteins?
d. What is the role of molecular genetic analysis?
2 Management
a. Does every child and adult with HS need folate supplementation?
b. What are the indications, timing and method of splenectomy?
c. What are the risks associated with splenectomy in HS?
What advice should be given with regard to postsplenectomy antibiotic prophylaxis and what is the necessity
and frequency of booster vaccinations?
d. How should gallstones be managed?
The definitions of the types of evidence and the grading
recommendations used in this guideline are listed in Appendix.
Search strategy
In addition to personal archives and consultation of standard
textbooks (together with relevant cited references therein), the
writing group searched Medline and Embase to identify relevant
literature including meta-analyses (none found), reviews and
original papers in any language, using the following key words
and combinations of them: hereditary spherocytosis; red cell
membrane; spectrin, ankyrin, band 3, spherocytes; haemolysis;
folate; folic acid; splenectomy; splenectomy in haematological
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
doi:10.1111/j.1365-2141.2004.05052.x
Guideline
conditions, cholecystectomy; cholecystostomy; laparoscopic;
gallstones; pneumococcal; vaccination; penicillin prophylaxis,
postsplenectomy sepsis, non-operative approaches to gall
stones. Only the abstracts were read of papers published in
languages other than English.
The biochemical basis of HS
The clinical expression of HS is relatively uniform within a
given family, but the disease severity varies considerably
between the different families.
which binds to beta-spectrin. In addition to binding protein
4Æ1 on its cytoplasmic N-terminal domain, band 3 interacts
with protein 4Æ2 to provide additional stability for the
cytoskeleton (Rybicki et al, 1996). Glycophorin C interacts
with protein p55 and protein 4Æ1, which in turn binds to beta
spectrin. Thus a deficiency of, or a dysfunction in, any one of
these membrane components within the vertical attachment
can weaken or destabilize the cytoskeleton, resulting in
abnormal red cell morphology and a shorter lifespan for
the affected red cells in the circulation.
Biochemical abnormalities associated with HS
Organization of red cell cytoskeleton
The abnormal morphology and shorter lifespan of the red cells
in HS are attributable to a deficiency or dysfunction of one of
the constituents of the red cell cytoskeleton (Delaunay et al,
1996; Tse & Lux, 1999) whose role is to maintain the shape,
deformability and elasticity of the red cell (Fig 1).
The red cell cytoskeleton is a spectrin-based network of
proteins located immediately on the cytoplasmic side of the
membrane lipid bilayer. Each spectrin protein (alpha and
beta) consists mainly of repeating units (106 amino acids
long) that fold into triple helices. The paired alpha/beta
spectrin heterodimers form head-to-head tetramers while the
other end of the spectrin heterodimer is associated with
protein 4Æ1 and actin (Fig 1). The vertical attachment of the
spectrin network to the lipid bilayer involves two transmembraneous proteins, band 3 and glycophorin C. Band 3,
existing as dimer and tetramers in situ in red cells, has
multiple protein-binding sites. It can interact with ankyrin,
The abnormal red cell morphology (resulting in shortened
cell survival) is due to a deficiency of, or a dysfunction in,
spectrin, ankyrin, band 3 and/or protein 4Æ2 (reviews
Delaunay et al, 1996; Tse & Lux, 1999). Severe haemolytic
anaemia is often associated with a greater reduction of the
affected membrane protein(s). There is an apparent correlation between clinical and protein phenotypes (Miraglia del
Giudice et al, 1994). HS red cells are found to have single or
combined protein deficiencies as determined by sodium
dodecyl sulphate-polyacrylamide gel electrophoresis (SDSPAGE). An analysis of 174 HS patients from 123 families in
the UK uncovered a larger proportion of spectrin-deficient
probands whose parents are apparently normal when compared with other protein deficiencies. Autosomal dominant
HS is often found to have primary mutations in the genes of
ankyrin, band 3 or b spectrin. Primary mutations in these
three genes can lead to a secondary protein deficiency
detected by SDS-PAGE (Table I).
Fig 1. Schematic presentation of the structural organization of red cell cytoskeleton (Bolton-Maggs, 2004). b Spectrin is the key component in that it
pairs with a spectrin to form a heterodimer, and it has binding sites for ankyrin and protein 4Æ1. The common protein defects are associated with
spectrin (a and/or b), ankyrin, band 3 protein and protein 4Æ2. Although protein 4Æ1 abnormality and spectrin mutations are usually associated with
hereditary elliptocytosis, a subset of hereditary spherocytosis was found to have b spectrin mutant proteins that had defective interaction with protein
4Æ1 (Goodman et al, 1982; Wolfe et al, 1982). This figure is a preprint from an article accepted for publication by Archives of Disease in Childhood,
and may not, save under the fair dealing of the provision of the Copyright Designs and Patents Act (1988), be reproduced without the consent of the
BMJ publishing group.
456
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
Table I. Probable primary causes for producing
the observed (or secondary) membrane protein
defects.
Observed protein deficiency by SDS-PAGE
Primary defect in protein or gene
Partial spectrin and protein 4Æ2 deficiency
Partial ankyrin and spectrin deficiency
Partial spectrin deficiency
Marked deficiency of spectrin (parents are normal)
Missing one haploid set of ANK1
Ankyrin gene mutation(s)
A variety of molecular defects
Severe ndHS due to low expression
allele inherited in trans to a second
Sp allele (i.e. HS)
Band 3 mRNA instability
Band 3 gene mutations resulting in
loss of protein 4Æ2 binding site
Partial band 3 deficiency
Partial protein 4Æ2 deficiency
Table II. Membrane molecules associated with erythrocyte cytoskeleton.
Protein
Band Mr
on gel (kD)
a Spectrin
b Spectrin
Ankyrin
Band 3 (AE1)
1
2
2Æ1
3
Protein 4Æ1
4Æ1
Protein 4Æ2
4Æ2
Glycophorin C GPC
240
220
210
90–100
80
72
32
Gene
SPTA1
SPTB
ANK1
AE1
(SLC4A1)
EPB41
EPB42
GYPC
Chromosomal Number
location
of exons
1q22-q23
14q23-q24Æ1
8p11Æ2
17q21-q22
52
32
42
20
1p36Æ2-p34
15q15-q21
2q14-q21
‡22
13
4
Molecular genetics of HS
The genes encoding membrane proteins of the red cell
cytoskeleton and their respective chromosomal locations are
known (Table II). Further information can be accessed from
websites such as http://www.kidscancer.net/membrane/ and
the Human Gene Mutation Database at http://www.hgmd.org
(Stenson et al, 2003).
Very different molecular mechanisms can generate the same
membrane protein deficiency in HS. Most of the reported
protein gene mutations in HS are ‘private’ or sporadic
occurrences, i.e. they are specific to one family or found in a
few families from different countries. Knowledge of the gene
mutation does not influence the clinical management of the
patient but analysis of mutant protein gene in family studies
can clarify one of the following conditions.
1 To identify the primary gene defect. Family study will yield
information on whether the abnormal protein in the
proband is encoded by a mutant candidate gene or due to
a secondary event resulting from a mutation in a different
protein gene (examples shown in Table I).
2 To confirm recessive or non-dominant (r/nd)-HS when
both parents are apparently normal. Different genetic
backgrounds can give rise to r/nd-HS, which have been
found in association with a spectrin, ankyrin and b spectrin
genes. Recent analysis of nd-HS in an Italian population
ª 2004 Blackwell Publishing Ltd, 126, 455–474
(Miraglia del Giudice et al, 2001) has shown that the
occurrence of de novo non-dominant mutations in HS
patients with normal parents is six times more common
than recessive mutations (level III evidence). The distribution of the nd-HS clinical phenotypes nearly overlapped
with the group of dominant HS.
The SpaLEPRA allele (LEPRA: Low Expression PRAgue) is
prevalent among nd-HS (Boivin et al, 1993; Wichterle et al,
1996; Dhermy et al, 2000). This allele remains silent when
inherited by a normal individual. However, mutations in the
spectrin gene (SPTA1), both homozygous and compound
heterozygous (co-inherited with a pathogenic HS allele),
result in severe HS with very low spectrin levels in the red
cells.
De novo mutations are mostly found in recessive HS
associated with the ankyrin gene (ANK1) (Eber et al, 1996)
and b spectrin (SPTB) (Miraglia del Giudice et al, 1998).
The transmission of these ankyrin mutations to subsequent
generations occurs in a dominant manner. A recurrent
frameshift mutation, Ankyrin Florianópolis, has been found
in three unrelated probands with severe dominant HS from
different genetic backgrounds (Gallagher et al, 2000).
In the case of band 3 gene defect, additive effects of two
unequally expressed AE1 mutant alleles can aggravate the
clinical features of an affected individual (Alloisio et al,
1996, 1997; Bracher et al, 2001). The mutations associated
with band 3 are heterogeneous with no apparent founder
effect.
Clinical features and diagnosis
Hereditary spherocytosis is a heterogeneous group of disorders
with regard to clinical severity, protein defects and mode of
inheritance. The clinical severity of HS varies from symptomfree carrier to severe haemolysis. Mild HS can be difficult to
identify because individuals may have a normal haemoglobin
and bilirubin. The presence of spherocytes and a reticulocytosis will support the diagnosis. If there are no spherocytes seen
on the film, no abnormalities in the red cell indices and the
reticulocyte count is normal, a ‘carrier’ state cannot be
excluded, but the individual is unlikely to have any clinical
457
Guideline
sequelae. Occasionally mild HS can be exacerbated by illnesses
that cause splenomegaly, such as infectious mononucleosis
(Gehlbach & Cooper, 1970).
Age at diagnosis
Although the diagnosis of HS is often made in childhood and
young adult life, it may be diagnosed at any time, even in the
seventh to ninth decades of life (Friedman et al, 1988). It is
perhaps surprising that the diagnosis of HS is not always
considered as a cause of gallstones and splenomegaly in adult
life, and that even in recent years there may be a considerable
interval between the onset of symptoms and diagnosis
(Reinhart et al, 1994). Asymptomatic HS may also be uncovered (usually in childhood) after an aplastic crisis, particularly
when associated with Parvovirus B19 (Summerfield & Wyatt,
1985; Lefrere et al, 1986; Cohen et al, 1991), or influenza
infection (Jensson et al, 1977). Hydrops fetalis or stillbirth can
occur when a fetus has inherited defects in membrane proteins
from both parents (Whitfield et al, 1991; Gallagher et al,
1995). Very mild HS may be first detected during investigation
of asymptomatic relatives of a known case or during pregnancy
(Kohler et al, 1960; Pajor et al, 1993).
Clinical features
The diagnosis of HS is generally straightforward. The patient,
usually a child, has a variable degree of anaemia, jaundice and
splenomegaly. It is important to ask for a family history of
jaundice and/or splenectomy, as families may not realize that
the cause involves their red blood cells.
Neonatal jaundice is common and may be severe enough to
require exchange transfusion, but is not clearly related to the
subsequent severity of HS. The mechanism for this is not
clearly understood. It should be noted that the diagnosis of HS
in neonates may be difficult (Schroter & Kahsnitz, 1983). The
film appearances may not be typical and the osmotic fragility
test is unreliable. As potential parents, adult patients need to be
advised of the risk of neonatal jaundice, as occasionally
exchange transfusion has been required for severe hyperbilir-
ubinaemia. There may be an interval of several days before
bilirubin reaches its peak, so infants born to known parents
with HS should be carefully monitored over several days.
Co-inheritance of other haematological disorders, such as beta
thalassaemia trait or SC disease, can lead to confusion in the
diagnosis and variable clinical effects (Gallagher et al, 1998).
Iron, folate or vitamin B12 deficiencies can mask the laboratory features. Obstructive jaundice alters the lipid composition of the red cell membrane, masking the film appearances
and reducing the haemolysis.
Family history
Most cases (75%) will have a family history of HS. In clear-cut
cases [typical history and physical findings, spherocytes on the
blood film with a reticulocytosis and negative direct antiglobulin test (DAT)] there is no indication to perform further
special laboratory tests to arrive at the diagnosis. In the absence
of a family history the most important differential diagnosis is
autoimmune haemolysis. This is rare in children, but can
follow a viral infection and is usually transient. Usually (but
not always) this can be excluded by a negative DAT (see
below). Other membrane abnormalities must be considered if
the morphology is atypical, and in these, further investigations
may be required (see below).
Consideration of individual clinical features
Anaemia
The severity of anaemia is closely associated with the severity
of haemolysis and spleen size. Many individuals have compensated haemolysis and have a normal haemoglobin level
with a reticulocytosis. During pregnancy some non-splenectomized HS patients may present with sufficiently severe
anaemia to warrant blood transfusion.
Symptoms of HS may appear in the perinatal period:
jaundice is common in the first 2 d of life. Some neonates with
HS may be transfusion-dependent due to their inability to
mount an adequate erythropoietic response in the first year of
Table III. Classification of spherocytosis and indications for splenectomy (modified from Eber, S.W., Armbrust, R. & Schroter, W. Variable clinical
severity of hereditary spherocytosis: relation to erythrocytic spectrin concentration, osmotic fragility and autohemolysis. Journal of Pediatrics, 177,
409–411 (copyright 1990, with permission from Elsevier).
Classification
Trait
Mild
Moderate
Severe
Haemoglobin (g/dl)
Reticulocyte count %
Bilirubin (lmol/l)
Spectrin* per erythrocyte
(% of normal)
Splenectomy
Normal
Normal (<3%)
<17
100
11–15
3–6
17–34
80–100
8–12
>6
>34
50–80
6–8
>10
>51
40–60
Not required
Usually not necessary
during childhood
and adolescence
Necessary during school
age before puberty
Necessary – delay until
6 years if possible
*Data on spectrin content are provided for interest; it is not necessary to measure this.
458
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
life (Delhommeau et al, 2000) (level III evidence). Continued
transfusion dependence is unusual and it is important to avoid
repeated transfusion where possible. Recent evidence suggests
that erythropoietin may be of benefit in reducing or avoiding
transfusion, and can usually be stopped by the age of 9 months
(Tchernia et al, 2000). Many older children with Hb levels of
5–6 g/dl do not require transfusion. Children who require one
or two transfusions early in life frequently become transfusion
independent.
Clinical features
Splenomegaly almost always
Laboratory red cell indices (flHb, flMCV, ›MCHC, ›% hyperdense
cells, ›RDW, ›reticulocyte count)
Blood film
Abnormal morphology – spherocytes
Direct antiglobulin test
Negative
Evidence of haemolysis
Raised bilirubin; reticulocytosis
Splenomegaly
MCV, mean cell volume; MCHC, mean cell Hb concentration; RDW,
red cell distribution width.
Most children and adults with HS have mild to moderate
enlargement of the spleen, but other than assisting in the
diagnosis, this is of little clinical significance. The size of the
spleen per se is not an indication for splenectomy. There is no
evidence from the literature that splenic rupture is commoner
than in the normal population.
Associated medical conditions may rarely be associated with
HS, particularly various neuromuscular conditions (Gallagher
et al, 1998), psychomotor retardation associated with abnormalities in chromosome 8 (also the location of the ankyrin
gene) (Chilcote et al, 1987; Cohen et al, 1991) and eye disease
manifested by angioid streaks (McLane et al, 1984).
Disease severity
Hereditary spherocytosis patients are usefully clinically classified
as ‘mild’, ‘moderate’ or ‘severe’ (for criteria see Table III), and
this correlates with the spectrin content detected in the red cell
membranes – more severe clinical disease being associated with
lower spectrin content. Furthermore, it predicts the clinical
behaviour and the need for and response to splenectomy (Agre
et al, 1986; Eber et al, 1990) (Table III – level III evidence).
Clinical manifestations of the disorder in a majority are
abrogated by splenectomy. However, those with severe HS,
most notably where the red cells have £50% spectrin content,
remain anaemic postsplenectomy (Agre et al, 1986), and some
of them may need blood transfusion during an infection.
Recently it has been shown that the nature of the membrane
defect (whether predominantly spectrin/ankyrin, or band 3)
influences the response to splenectomy (Reliene et al, 2002). The
assessment of severity should be made when the patient is well,
preferably with a number of baseline tests over a period of time,
rather than relying on the results and picture which first drew the
patient to medical attention when perhaps unwell and decompensated, leading to a false assessment of the severity of HS.
Recommendation
Patients with HS should be graded by their severity of
disease (baseline Hb, reticulocyte count, jaundice, level of
activity). This predicts clinical course and the need for
splenectomy. See Table III (level III evidence, grade B
recommendation).
ª 2004 Blackwell Publishing Ltd, 126, 455–474
Table IV. Diagnostic parameters for hereditary spherocytosis.
Parameter
Features
Laboratory investigation
The laboratory diagnosis is usually straightforward and is based
upon a combination of clinical history, family history, physical
examination (splenomegaly, jaundice) and laboratory data (full
blood count, especially red cell indices and morphology, and
reticulocyte count) (Table IV). Other causes of haemolytic
anaemia should be excluded, particularly autoimmune haemolytic anaemia caused by warm (IgG) or cold (IgM) autoantibodies, i.e. positive DAT. Autoimmune haemolysis (AIHA) can
usually be excluded by a negative DAT. AIHA is uncommon in
children and would not show the strong family history
commonly present in HS. There can be difficulty in interpreting
DAT-negative AIHA in an individual with no family history,
particularly an adult. In some cases the density of attached
autoantibody may be too low for detection by DAT. This may be
resolved by flow cytometric assessment of red cell Ig density
(Kerr et al, 2000). Polyspecific DAT reagents may also fail to
detect some autoantibodies, particularly IgA autoantibodies, as
anti-IgA is not usually incorporated into the reagent. Application of specific anti G, A, M and complement DAT reagents may
resolve some cases (Bardill et al, 2003).
In neonates haemolysis caused by irregular maternal antibodies must be excluded, at least by a negative DAT. The
diagnosis of HS may be difficult in the neonatal period. If the
baby is well, testing can be postponed until the child is at least
6 months of age or older when the morphology may be less
confusing. The eosin-5-maleimide (EMA) binding dye test (see
below) will be positive irrespective of morphology. Other
membrane abnormalities must be considered if the morphology
is atypical, and in these, further investigations may be required.
Red cell morphology
The appearance of spherocytes in HS is attributed to a spleenmediated loss of microvesicles from discoidal red cells. In
addition to spherocytes and/or microspherocytes, HS red cells
with diverse morphology have also been noted. Acanthocytes
or spiculated red cells (8–15%) were found in patients with b
spectrin gene mutations (Wolfe et al, 1982; Hassoun et al,
459
Guideline
Table V. Elevated automated red cell parameters in the prediction of hereditary spherocytosis and its severity.
Parameters
Remarks
Reference
Combined MCHC and RDW (or HDW)
Combined % hyperdense* cells and RDW
Significantly higher in HS
Excellent diagnostic efficacy and
indication of HS severity
The best indicator of HS severity
A discriminating feature of the HS phenotype
Pati et al (1989), Michaels et al (1997)§
Pilar Ricard & Gilanz (1996)
% Microcytes
Combined MCHC and % hyperdense cells
Cynober et al (1996)
Cynober et al (1996)
HDW, Hb distribution width.
*Hyperdense cells are also present in Hb SC disease, Hb CC disease and xerocytosis.
Physiological microcytosis occurs in children (Dallman & Siimes, 1979).
Dual angle laser light scattering haematology analyser.
§Electronic aperture impedance haematology analyser.
1997), and pincered (or mushroom-shaped) cells have been
found in unsplenectomized HS with band 3 deficiency (Jarolim
et al, 1996; Dhermy et al, 1997). In a case of severe haemolytic
HS with a combined deficiency of spectrin and ankyrin, red
cells with an irregular contour (resembling red cells of
hereditary pyropokilocytosis) were seen (Palek & Sahr,
1992). Ovalostomatocytosis is a consistent feature for Japanese
HS patients whose red cells have a protein 4Æ2 deficiency
resulting from a protein 4Æ2 gene mutation in nt
142GCT fi ACT (Yawata et al, 1994).
It is important to differentiate HS from hereditary stomatocytosis and related disorders with abnormal permeability of
sodium and potassium ions. These disorders are rare; the
morphology may not be typical, particularly where films are
not freshly made. Such families have been labelled as ‘atypical’
HS. Splenectomy is not effective and is associated with a high
risk of thrombosis (Stewart et al, 1996; Stewart & Turner,
1999; level III evidence).
An artefact showing ‘macrospherocytosis’ on a blood film
can be produced as a result of cold storage of blood samples
from patients with cryohydrocytosis, which is a variant form of
hereditary stomatocytosis (Haines et al, 2001). This blood
picture may inadvertently suggest a diagnosis of ‘atypical HS’.
Although hereditary stomatocytosis and related disorders with
abnormal permeability of sodium and potassium ions are rare,
their differentiation from HS is essential because HS red cells
have also been found to have abnormal cation membrane
permeability, unrelated to a specific membrane protein defect
(De Franceschi et al, 1997). Moreover, splenectomy is not
effective for the dehydrated and overhydrated forms of
hereditary stomatocytosis, and is associated with a high risk
of thrombosis (Stewart et al, 1996; Stewart & Turner, 1999).
Red cell indices
Haematology analysers using the principle of flow cytometry
(i.e. dual angle laser light scattering method) produce a more
accurate determination of red cell volume (MCV) and
haemoglobin concentration (MCHC) than those blood cell
analysers using the principle of electrical aperture impedance
(Mohandas et al, 1986). These automated red cell parameters
460
can be used to predict or identify HS with a typical clinical
presentation during the routine full blood count (Tables IV
and V) without requiring additional laboratory tests (such as
osmotic fragility or the EMA binding test) to confirm the
diagnosis. These parameters are more pronounced in splenectomized than in non-splenectomized HS patients.
Additional tests for the diagnosis of HS
Additional testing for confirmation of HS is indicated when
the diagnostic criteria are not met, and other causes of
haemolysis have been excluded. For instance, the film appearances are atypical, there is no clear pattern of inheritance, or
the proband has an on-going mild haemolytic process, albeit
with an apparently normal full blood count result. The
diagnosis may be particularly difficult in the neonatal period
because spherocytes are commonly seen in neonatal blood
films (Schroter & Kahsnitz, 1983).
Screening tests. All the laboratory tests currently in use can
detect typical HS (Table VI). The acid glycerol lysis test has a
higher detection rate in asymptomatic relatives of known
affected individuals than the osmotic fragility test. The
drawback of both tests is an apparent lack of specificity
under certain circumstances. In addition to detecting HS, they
give positive results with a wide spectrum of clinical disorders
or red cells in which abnormalities are unrelated to structural
protein defects (Table VI). For instance, the osmotic fragility
test cannot differentiate between causes of spherocytosis
(immune versus non-immune).
A normal osmotic fragility result does not exclude the
diagnosis of HS and may occur in 10–20% of cases of HS (Dacie
et al, 1991). The test may also be normal in the presence of iron
deficiency, obstructive jaundice, and in the recovery phase from
an aplastic crisis when the reticulocyte count is increased
(Korones & Pearson, 1989). Cell dehydration occurring in the
spherocytes of a patient with HS can be one of the causes of
normal osmotic fragility results for non-splenectomized HS
patients (Cynober et al, 1996). Furthermore, a positive osmotic
fragility result can be obtained in patients with hereditary
elliptocytosis (HE) and haemolysis.
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
Table VI. Screening tests for the diagnosis of hereditary spherocytosis.
Test
Method
Evaluation (sensitivity, specificity, precautions)
Osmotic fragility (OF) test*
(Parpart et al, 1947)
Heparinized whole blood (50 ll) per tube. Measure
absorbance at 540 nm for fresh blood and after
24 h incubation. Plot a graph of % haemolysis
versus NaCl concentration
Affected by elevated reticulocyte count: increased
OF in 66% of non-splenectomized HS patients.
Also increased OF found in immune-mediated
and other haemolytic conditions
Acidified glycerol lysis test
(AGLT) (Zanella et al, 1980)
The Pink test (Bucx et al, 1988)
is a modified AGLT
EDTA (20 ll )whole blood per tube. Measure the
time taken for absorbance of red cell suspension at
625 nm in glycerol to fall to half of its original
value before glycerol addition (AGLT50)
Also detects autoimmune haemolytic anaemia,
hereditary persistence of fetal haemoglobin,
pyruvate kinase deficiency, severe glucose6-phosphate dehydrogenase deficiency,
pregnant women (one-third), chronic renal
failure on dialysis (some), and myelodysplastic
syndrome. Reagent preparation: special attention
to pH and osmolality
Osmotic gradient ektacytometry
(Clark et al, 1983)
A laser diffraction viscometer that measures red cell
deformability at constant shear stress as a
continuous function of suspending osmolality
(hypotonic to hypertonic)
Distinct deformability curves for red cells from
patients with HS, hereditary elliptocytosis,
hereditary pyropoikilocytosis, stomatocytosis
and sickle disease (Mohandas et al, 1980)
Hypertonic cryohaemolysis test
(Streichman & Gescheidt, 1998)
Packed red cells (50 ll) per tube. % cryohaemolysis
at 540 nm after transfer of red cells from 37C to
0C for 10 min
Positive results for HS, some CDAII and
Melanesian elliptocytosis
Eosin-5-maleimide (EMA) binding§
(King et al, 2000)
Packed red cells (5 ll; EDTA sample) per tube.
Reduced fluorescence (green) intensity of
EMA-labelled red cells by flow cytometry
Distinct histograms for red cells of HS
(sensitivity of 92Æ7% for HS and a specificity of
99Æ1%) and HPP. Reduced fluorescence with
CDAII, cryohydrocytosis, SAO. Reagent:
solution of EMA (light-sensitive) must be
stored at )20C (max. 3 months)
*The shift of OF curve reflects the red cell response to hypotonic medium due to reduced cell volume and surface area.
The red cell lipid composition regulates the membrane permeability to glycerol. The added glycerol slows down the rate of water entry into the red
cells. This allows the measurement of time taken for cell lysis to occur.
There is no ektacytometer in the UK. Measurement of red cell membrane resistance to fragmentation under shear stress.
§EMA binds to band 3, Rh blood group proteins, Rh glycoprotein and CD47 on membrane molecules in intact red cells (King et al, 2004).
The cryohaemolysis test, the osmotic gradient ektacytometry
and the EMA binding test have a higher predictive value in the
diagnosis of HS because there have been no reports of positive
results in immune or non-membrane-associated disorders.
However, these tests are not specific, and can also detect red
cells with rare membrane disorders, such as aberrant band 3
proteins [e.g. CDAII, Southeast Asian ovalocytosis (SAO)], a
change in intracellular viscosity (e.g. sickle cells), and temperature-sensitive monovalent cation transport (e.g. cryohydrocytosis) (Table VI). In mild or atypical cases, difficulty in
interpretation is likely to occur with results that fall between
normal and typical HS reference ranges. Therefore, ektacytometry has an advantage in that results are plotted as a deformability
curve, which has a distinct shape for each type of abnormal red
cells tested (Mohandas et al, 1980). In EMA binding by flow
cytometry, hereditary pyropoikilocytosis (HPP, MCV < 60 fl)
and HE can be differentiated from HS, based on the graded
reduction in fluorescence intensity for HPP (the lowest)<
HS < HE £ normal controls (King et al, 2000, 2003). Although
similar fluorescence readings to the red cells of HS are obtained
with some rare red cell disorders – CDAII, SAO (also known as
ª 2004 Blackwell Publishing Ltd, 126, 455–474
Melanesian elliptocytosis) and cryohydrocytosis – these can be
distinguished from HS on the basis of their distinct clinical
features (Table VII). Other atypical cases with spherocytosis can
produce anomalous histograms, which are identified by overlaying the test histogram with those of normal and HS controls
(King et al, 2003). Normal fluorescence results are obtained for
individuals presenting a reticulocytosis or an autoimmune
haemolytic anaemia (King et al, 2000).
In conclusion, when selecting an appropriate test for
determining mild or atypical HS condition in a patient, the
sensitivity and specificity of the test for HS, the complexity of
the protocol, and the total cost of instrument(s) and its
maintenance should be taken into consideration. As HS is a
structural defect of the red cell cytoskeleton, the test result is
expected to give a clear distinction between a membraneassociated and a non-membrane-related red cell disorder.
SDS-PAGE. Identifying a deficiency in a membrane protein
associated with erythrocyte cytoskeleton confirms the
diagnosis of HS. Quantitation of membrane proteins by
SDS-PAGE is, however, not necessary for the majority of HS
461
Guideline
Table VII. Features of South-East Asian ovalocytosis (SAO), congenital dyserythropoietic anaemia type II (CDAII), and hereditary stomatocytosis
(HSt.) and related disorders.
Disorder
Mode of
transmission
Red cell indices and
morphology
Clinical features
SAO*
Autosomal
dominant
Asymptomatic of any haemolysis
Rigid and hyperstable.
25% or more ovalocytes.
Some SAO cells show a
‘slit’ in the cells
CDAII (HEMPAS)
Autosomal
recessive
Mild and intermittent jaundice,
and splenomegaly. Hepatomegaly
and gallstones are less frequent.
Anaemia, liver cirrhosis and
secondary tissue siderosis
Overhydrated HSt.
(hydrocytosis)
Dominant
Dehydrated HSt.
(xerocytosis)
Dominant
Familial pseudohyperkalaemia
Dominant
Frequency 1 in 1 million births.
Indistinguishable from HS
(atypical HS)
Frequency 1 in 10 000; heterogeneous:
combined with pseudohyperkalaemia
and/or perinatal oedema
Non-haemolytic (normal haematology
without jaundice)
Bi- and multinucleated
erythroblasts (usually
15–30%). Mild to
moderate Hb level, normal
to mild elevated MCV.
Low reticulocyte count
Stomatocytes (some have
30%) decreased MCHC,
macrocytosis
Lack deformability: some
stomatocytes and target
cells; increased MCHC
Normal
Cryohydrocytosis
Dominant
Rare, mild stomatocytic haemolytic
state with hyperbilirubinaemia,
gallstones
Red cell lysis when stored
at 4C (increased
intracellular Na+)
Laboratory test
(a) SDS-PAGE detects a slower
migrating Band 3 Memphis
(Lys56 fi Glu) protein
(b) Detect deletion of
27 nts in band 3 gene
A more compact and faster
migrating band 3 (due to
incomplete N-glycosylation)
by SDS-PAGE. Confirmation of
CDA by electron microscopy of
bone marrow sample
Stomatin absent (band 7Æ2 in
SDS-PAGE); intracellular
[Na+] > 60 mmol/l red cells
Normal stomatin. Measure
intracellular [Na+] and [K+]
Normal or near normal
intracellular Na+, net loss of [K+]
when stored at room temperature;
apparently normal cation
fluxes at 37C
Measure intracellular [Na+] and [K+]
*Also known as Melanesian hereditary elliptocytosis. Found mainly in Malaysia, Indonesia, Papua New Guinea, the Philippines. Occasionally in South
African (White), one Afro-American family.
Fig 2. Flow chart for the diagnosis of hereditary
spherocytosis (modified from Iolascon et al
1998).
462
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
cases because a definitive diagnosis can be made on the basis of
red cell indices, clinical/family history and positive result from
a screening test. Nevertheless, SDS-PAGE can be useful in HS
cases where the index member has a clinical condition
inconsistent with that of the affected parent(s) and siblings
(see Biochemical abnormalities associated with HS).
Information on genetic analysis. There are currently no
facilities for the determination of molecular defects in the
protein genes associated with HS in the UK.
The DNA-based methods in the diagnosis of common
inherited diseases are applicable to the detection of erythrocyte
membrane protein gene mutations. The rapid screening based
on single-strand conformation polymorphism can be carried
out to detect mutations in the coding region of cDNA of all the
HS-associated genes. Linkage analysis has been successfully used
in the search for candidate genes in dominant HS (Costa et al,
1990; Garbarz et al, 1996, 1998). Determination of the (AC)n
microsatellite-length polymorphism at the 3¢ end of the ankyrin
gene provides an indicator for ankyrin abnormality (e.g. an
absence of reticulocyte ankyrin mRNA) (Jarolim et al, 1995;
Eber et al, 1996).
Recommendations
A suggested diagnostic pathway is shown in Fig 2 –
modified from Iolascon et al (1998).
1 Newly diagnosed patients with a family history of HS,
typical clinical features (splenomegaly) and laboratory
investigations (spherocytes, raised MCHC, increase in
reticulocytes) do not require any additional tests (level
III evidence, grade B recommendation).
2 If the diagnosis is equivocal, for example, where there are a
few spherocytes on the film but no other laboratory,
clinical or family evidence, a screening test with high
predictive value for HS is helpful (level IV evidence, grade
C recommendation). The recommended screening tests
are the cryohaemolysis test and the EMA binding test
(level IIa/III evidence, grade B recommendation). The
high predictive value of both techniques for the diagnosis
of HS can be improved further when the results are used in
conjunction with clinical information and red cell indices.
The availability of a flow cytometer may be the deciding
factor for choosing the EMA binding method. In view of
the probable inherent differences in flow cytometers, each
laboratory must set up its own cut-off value for distinguishing between normal and HS red cells (mean ± 2SD),
and normal reference range (King et al, 2000) (level IIa
evidence, grade B recommendation). Osmotic fragility
may still be performed purely as a supportive investigation, but its limitations must be recognized. It should be
possible for all laboratories to make arrangements for the
EMA dye test to be provided by a collaborating service.
ª 2004 Blackwell Publishing Ltd, 126, 455–474
3 Confirmation of the diagnosis may be necessary in
selected cases if the screening tests produce an equivocal
or borderline result. Gel electrophoresis analysis of
erythrocyte membranes is the method of choice (level III
evidence, grade B recommendation). This technique is
useful for determining the extent of membrane deficiency for the patient. The main drawback is a lack of
sensitivity to very mild or asymptomatic ‘carrier’ HS.
The use of SDS-PAGE is recommended:
a. When the clinical phenotype is more severe than
predicted from the red cell morphology.
b. When the red cell morphology is more severe than
predicted from parental blood films where one parent
is known to have HS.
c. If the diagnosis is not clear prior to splenectomy,
when a patient might have an abnormality in monovalent cation permeability. Where the morphology is
typical there should be no doubt. In more subtle cases
(when MCV > 100 fl) clarification may be essential
(Delaunay et al, 1999). Splenectomy may not be
appropriate for the clinical management of these
patients (Stewart et al, 1996; Stewart & Turner, 1999)
(level IIa evidence, grade B recommendation).
4 Diagnosis of HS does not require further investigation
by molecular analysis of the affected genes (level III
evidence, grade B recommendation).
Clinical management of individuals with HS:
medical
Indications for folate therapy
Folate supplements have traditionally been prescribed to all
individuals with chronic haemolysis. Is this required, particularly in mild cases of haemolysis? Megaloblastic anaemia has
been reported in 11 cases of HS, including two children (level
III evidence). Review of the literature led to the conclusion that
the evidence ‘although highly fragmentary, therefore suggests
that folic acid deficiency has been responsible in the majority
of cases’ (Lindenbaum & Klipstein, 1963). Folate replacement
is probably only required as a routine for children with severe
haemolysis, (Rosenblatt & Hoffbrand, 1999) and in pregnancy,
whatever the severity of the HS. Various regimens are
suggested – 1 mg/d, 5 mg once a week, 100 lg/kg/d. The
daily requirement in normal adults is only about 200 lg/d and,
for children the World Health Organization (WHO) recommended intake is 3Æ3–3Æ6 lg/kg/d. Consideration should be
given to the socio-economic environment of the child (folate
deficiency is more likely to occur in association with poverty),
and their diet, in addition to the severity of the HS, before
committing a child to lifelong medication. Folates are wide-
463
Guideline
spread in food, especially in fresh fruit and vegetables (which
must not be overcooked). Breakfast cereals are also an
important source in children aged 1–4 years. The National
Diet and Nutrition Survey in the UK showed that at all ages the
average daily intake in 1995 was well above the reference
nutrient intake, being 143% for children over 4 years, and
184% in children under 4 years, suggesting that folate
supplementation is unnecessary in mild HS.
Recommendation
Folate therapy is recommended in severe and moderate HS,
but is probably not necessary in mild HS. (level III and IV
evidence, grade C recommendation). A reasonable daily
dose would be 2Æ5 mg/d up to the age of 5 years, and 5 mg/
d thereafter.
Routine observation and frequency of blood tests
Children. Once a diagnosis of HS has been established a child
should be regularly followed up; for most children (those with
mild and moderate HS) an annual visit is sufficient once the
baseline has been established, and in the absence of symptoms
a blood count at every visit is unnecessary. If possible, the same
team should see the child and parents at each annual visit. This
is particularly important in the years leading up to the decision
for or against splenectomy. At each visit, general health,
growth, spleen size and exercise tolerance should be recorded.
A locally written explanatory booklet should be provided. If
this is not available the families should be given advice about
sources of information about HS. It is important to provide
information on parvovirus B19, as an aplastic crisis is
potentially dangerous and is certainly the most frightening
complication for most families. Children who are not receiving
regular folate supplements may have their folate status checked
to ensure levels are sufficient (see above). Ultrasound
examination for biliary stones may be performed from the
age of 5 years at convenient intervals – probably every 3–
5 years is sufficient in the absence of symptoms. Severely
affected children will require haematological supervision
during intercurrent viral infections, when they may be
decompensated. Children who are regularly transfused in the
early years (a very small group) should receive genotyped
leucodepleted blood in order to reduce the risk of reactions
and alloimmunization. (level III and IV evidence, grade C
recommendations).
Adults. The value of an annual visit in those with mild disease
is debatable. The disruption to their lives outweighs any
benefit. Providing the patient and general practitioner are
464
aware of the disorder it is unnecessarily disruptive to insist on
regular review. There is however perhaps some benefit in
regular annual review of those with moderate disease (and to
monitor by ultrasound examination for the development of
gallstones) or those who have had splenectomy, if only to
discuss current views about postsplenectomy management (revaccination, antibiotic prophylaxis). It is also worth checking
fasting iron status annually (transferrin saturation) in nonsplenectomized cases. Chronic anaemia will enhance iron
absorption, and co-inheritance of the haemochromatosis gene
(C282Y) even in the heterozygous state can lead to severe iron
overload. (grade C recommendations, based on level IV
evidence.)
Clinical management of individuals with HS:
surgery
This section will address the following questions:
1
2
3
4
5
In which patients is splenectomy indicated?
When should splenectomy be performed?
What approach to splenectomy should be used?
Should a concurrent cholecystectomy be performed?
Postsplenectomy thrombosis – what is the risk?
In which patients is splenectomy indicated?
Splenectomy is very effective in reducing haemolysis, leading
to a significant prolongation of the red cell lifespan, although
not necessarily to normal (Chapman & McDonald, 1968;
Baird et al, 1971) (level III evidence). The clinical manifestations and complications (anaemia and gallstones) are much
reduced in severe HS and abolished in milder cases, but at
the price of an increased risk of life-threatening sepsis
from encapsulated organisms, particularly Streptococcus
pneumoniae.
Patients should be selected for splenectomy on the basis of
their clinical symptoms (Table III) and presence of complications such as gallstones, not simply on the basis of the
diagnosis alone (level III evidence, grade B recommendation).
Splenectomy should be performed in children with severe HS,
considered in those who have moderate disease, and should
probably not be performed in those with mild disease. A
careful history is important in those with moderate to mild
disease to establish if there is evidence suggesting reduced
exercise tolerance; the metabolic burden of increased marrow
turnover may be considerable. Where there is a family history,
the benefit of splenectomy in other individuals may help to
determine whether to proceed.
An analysis of decision-making for mild HS, based on the
available literature and computer modelling, suggests that
splenectomy is of no benefit in the absence of gallstones
(Marchetti et al, 1998). Children or young adults with mild
HS who also have gallstones are likely to benefit from
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
combined splenectomy and cholecystectomy in terms of life
expectancy. However, this model assumed 100% compliance
with penicillin prophylaxis, and if only 10% of splenectomized patients are non-compliant (which is likely) then
splenectomy is considered too risky for patients with mild
HS at any age. A difficult group are those with visible
jaundice but no stones, and who want splenectomy for
cosmetic reasons, but who have well compensated HS by
other criteria. However, these children often have a significant
reticulocytosis and splenectomy may be justified on the basis
that they are therefore likely to have symptomatic gallstones
later in life.
Infection risks and their management. The risk of
postsplenectomy sepsis is not completely eliminated by the
currently recommended preoperative vaccinations and
postsplenectomy antibiotic prophylaxis (BCSH, 1996; Davies
et al, 2002). There are numerous studies documenting an
increased mortality from sepsis; particularly notable was a
study of more than 1400 children in the US (Eraklis & Filler,
1972) and an early UK study of more than 800 children
demonstrating an age-related risk (Walker, 1976), which was
highest in the younger children, <2 years of age, and early after
splenectomy. A study of 456 children who had splenectomies
in Denmark between 1969 and 1978 also demonstrated an
incidence of pneumococcal bacteraemia and meningitis in the
splenectomized children that was 284 times that of nonsplenectomized children (Pedersen, 1983). Although the risk is
maximal in young children and early after splenectomy, it
persists for life (Ein et al, 1977; Evans, 1985; Holdsworth et al,
1991; Deodhar et al, 1993; Eber et al, 1999). The risk has been
reduced by deferring splenectomy until a child is over 6 years
of age, together with vaccination and antibiotic prophylaxis
(Konradsen & Henrichsen, 1991). The small but definite
remaining risk of sepsis must be clearly explained to the
parents and the child, and the indications for splenectomy
must be clear. Many adults who themselves may have had
splenectomy in the past may be unaware of the risks.
Conversely, adults who have received vaccinations and
antibiotics may perceive that there is no risk of infection and
also need appropriate education.
Presplenectomy vaccination and postsplenectomy followup. Patients should be vaccinated according to national
guidelines (BCSH, 1996; Davies et al, 2002). Fatal infection
may occur with a serotype not covered by the vaccine, or there
may not have been an adequate response to some of the 23
serotypes in the face of an adequate global response in terms of
functional antibody testing. Although repeat pneumococcal
vaccination is recommended at 5-year intervals, there is no
clear evidence in the literature. Some data exist demonstrating
that antibody levels may fall off quite quickly (Spickett et al,
1999). Good longitudinal follow-up studies of splenectomized
or hyposplenic patients to look at some of these issues are
ª 2004 Blackwell Publishing Ltd, 126, 455–474
badly needed (Reid, 1994). The recent update of the BCSH
guidelines (Davies et al, 2002) did not recommend any
changes to the current policy with the introduction of new
pneumococcal vaccines, but this needs regular review as more
evidence is gathered.
Antibiotic prophylaxis. Penicillin prophylaxis is recommended
for life postoperatively, but there is little published evidence
of efficacy (Reid, 1994). The guidelines indicate that
antibiotic prophylaxis should be for life, although the
evidence base for this is not strong (BCSH, 1996; Davies
et al, 2002). Urgent medical attention (even hospital
admission) should be sought in the event of a febrile
illness – however it is difficult to translate this into a
practical reality with children who are often febrile with
minor viral illnesses.
Despite recommendations for vaccination and antibiotic
prophylaxis (BCSH, 1996), ignorance persists in affected
families, and compliance with antibiotic advice is poor
(Waghorn & Mayon-White, 1997; Keenan et al, 1999; Spickett
et al, 1999). Although preventive measures are very successful
(Konradsen & Henrichsen, 1991) it is clear that they do not
completely eliminate the risk. The risk of infection remains
highest in the youngest patients but was reported to be reduced
by 47%, and the mortality by 88% in one study (Jugenburg
et al, 1999). Fatal pneumococcal infections have occurred
despite both vaccination and compliance with penicillin
prophylaxis (Deodhar et al, 1993; Spickett et al, 1999). A
recent study references at least 19 severe postsplenectomy
infections, six fatal, in vaccinated patients (Eber et al, 1999)
(level III evidence).
Recommendations
1 Patients and parents should be informed about the
lifelong small risk of overwhelming sepsis after splenectomy, and provided with a splenectomy card.
2 National guidelines for management should be followed
(BCSH, 1996; Davies et al, 2002). The need for reimmunization and its frequency are unclear as are the
optimal duration of postsplenectomy antibiotic prophylaxis and choice of drug.
When should splenectomy be performed?
No published evidence of the optimum timing for any type of
resection exists. All the major texts essentially state that the
timing is one of clinical judgement and that it depends on the
severity of the patient’s symptoms as judged by the effects of
anaemia, or the need for transfusion, or symptoms related to
cholelithiasis. More severe clinical disease (transfusiondependent) will benefit from early splenectomy but the risks
must be balanced.
465
Guideline
Recommendations on timing of surgery in relation to
postsplenectomy sepsis
The risk of serious sepsis is highest in young children and
in the first few years postsplenectomy (level III evidence).
Splenectomy should therefore be carried out as late as
possible, preferably after the age of 6 years; before this age
it is only warranted in very severe transfusion-dependent
disease and even so, should not be carried out before the
age of 3 years (Gallagher et al, 1998) (grade B recommendation). It is convenient to perform surgery before children
enter secondary education and before puberty places an
additional burden on erythropoiesis, i.e. before the age of
12 years. The risk of serious sepsis persists for life (level III
evidence).
What approach should be used?
Laparotomy. The traditional approach to splenectomy has
been total, by laparotomy, through either an upper midline
incision or more usually a left sub-costal approach. The spleen
has been removed in its entirety and a careful search made for
any splenunculi (accessory splenic tissue) in the assumption
that any splenic tissue left behind may lead to a recurrence of
symptomatic anaemia. There are no studies that have
quantified this risk, although it has been reported
(Mackenzie et al, 1962).
Laparoscopic splenectomy. The traditional approaches are
being challenged by the advent of laparoscopy. There are no
published randomized trials comparing laparoscopic
splenectomy to traditional approaches, although there are a
number of small reviews outlining experience with
laparoscopic splenectomy, either on its own (Smith et al,
1994; Rescorla et al, 2002; Tanoue et al, 2002) or in
combination with cholecystectomy (Caprotti et al, 1999).
There are several other studies that detail the outcomes of
cholecystectomy alone (Rescorla & Grosfeld, 1992; Holcomb,
1998) (level III evidence). Each of these is a descriptive study
and compares the outcomes of an institution’s experience
against either historical data from the same institution or
reported results from other institutions. These latter results are
often from a different time period. Published results
demonstrate that the laparoscopic approach is both feasible
and safe (Danielson et al, 2000), although there is no good
evidence of benefit over the more traditional approach.
Laparoscopic splenectomy is more difficult in the presence of
significant splenic enlargement (Bagdasarian et al, 2000) so
that a preoperative assessment of splenic size by ultrasound is
recommended (Esposito et al, 1998). There is a potential for
shorter hospital stay, fewer problems with wound infections,
improved pain relief (Curran et al, 1998), although in practice
none of this has yet been demonstrated by randomized
466
controlled trials. In general, it is a feasible approach from
surgeons who have the necessary equipment, training and
experience to carry out these procedures safely. The spleen is
caught in a specimen bag introduced through the port site, and
fragmented inside this so that the contents can be removed
through the port. Cholecystectomy may be performed at the
same time (Farah et al, 1997). Whether the traditional or
laparoscopic route is chosen will be determined by local
practice and expertise.
Partial splenectomy. Concern over the possible consequences
of sepsis has led some groups to investigate whether it is
necessary to remove the whole spleen to control haemolysis.
Partial splenectomy can improve transfusion-dependent
infants with very severe HS while preserving some splenic
function as protection against sepsis (level III evidence). A
number of reports have been published (Tchernia et al, 1993,
1997; Bader-Meunier et al, 2001). Forty patients were followed
for 1–14 years and it was shown that the majority remained
symptom-free, although three patients subsequently required
total splenectomy for further symptomatic anaemia (BaderMeunier et al, 2001), but only 17 of 40 patients had completed
at least 5 years follow up and a further five were lost to followup. It is too early to determine whether others will need total
splenectomy. The phagocytic function of the spleen was
sustained (demonstrated by technetium 99m scans) and
there were no reports of serious infection. It was not clear
whether the ability of the spleen to participate in humoral
defence was maintained, thus these patients were
recommended to continue with both immunizations and
anti-bacterial prophylaxis. Another group have also reported
favourable results in 16 children with HS followed for up to
6 years after partial splenectomy (Rice et al, 2003). Both of
these series are small, and have been followed for too short a
time to determine whether this procedure reduces the risks of
postsplenectomy sepsis. A further caution is that four of 18
patients developed new gallstones after subtotal splenectomy.
Should a concurrent cholecystectomy be performed?
In children with no evidence of cholelithiasis, there is no
indication to remove the gall bladder at the time of splenectomy.
Symptoms of cholelithiasis remain a prime reason for
carrying out a splenectomy in hereditary spherocytosis.
Stones themselves are reported to be present in 21–63% of
cases (Rutkow, 1981). Once the spleen is removed, individuals with HS do not develop pigment stones. In the absence
of stones, splenectomy alone is sufficient. A recent review of
17 patients who had undergone splenectomy without
cholecystectomy (gall stones excluded pre- or peri-operatively) under the age of 18 years demonstrated that none
developed any evidence of cholelithiasis (clinical or sub
clinical) over a mean follow-up of 15 years (Sandler et al,
1999).
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
A significant number of children with HS develop pigment
stones in the first decade of life. The risk is increased in
individuals who co-inherit Gilbert syndrome with HS – these
individuals have a fivefold increased risk of developing
gallstones (Miraglia del Giudice et al, 1999) (level III evidence). An Italian survey of 468 children with HS showed that
79 had gallstones, about half of them were diagnosed before
the age of 11 years (level III evidence) (Pinto et al, 1995). A
recent study of 103 unsplenectomized children has shown that
the pigmented stones seen in haemolytic diseases are radioopaque in 50%, but ultrasound has an accuracy of 96%
(Miraglia del Giudice et al, 1999). Regular ultrasound examination of the biliary system from the age of about 5 years may
identify individuals more likely to have troublesome symptoms later in life, and who may benefit from splenectomy
prior to puberty. It is not clear whether adults with mild HS
should be regularly screened by ultrasound for asymptomatic
gallstones, particularly as it is not clear whether these should
be removed. One study has shown chronic inflammatory
changes in the gall bladder mucosa in individuals with
asymptomatic stones (Csendes et al, 1998). However, there
are no good longitudinal studies of the outcome in children
with silent stones, if left alone. One study of 123 adults with
silent stones (not relative to haemolytic conditions) found that
only 15–20% developed symptoms over a prolonged follow-up
(Gracie & Ransohoff, 1982). In both adults and children,
regular follow-up should include assessment of possible
symptoms (typical and atypical) suggestive of gall bladder
disease. Individuals with symptomatic gallstones usually have a
cholecystectomy at the same time as the splenectomy; however
some surgeons prefer cholecystostomy (simple removal of
stones, leaving the gall bladder) in young children, as the risk
of stones is markedly reduced after splenectomy and there is
evidence that cholecystectomy may lead to alterations in bile
salt metabolism that predisposes to colon carcinoma later in
life, but the evidence for this is controversial (Robertson et al,
1988).
There are no randomized controlled trials comparing
different approaches to management but, in general, there is
a consensus for removing the gall bladder in the presence of
stones, if there have been symptoms of gall bladder disease
(biliary colic, cholecystitis). The consensus is less clear about
the appropriate action when stones are asymptomatic, and an
incidental finding. Although some authors have suggested that
all patients with stones, even the asymptomatic ones, should
undergo cholecystectomy (Gotz et al, 1977), there is enough
evidence to suggest that removal of stones may alone be
sufficient after splenectomy. Robertson et al (1988) reported
on the clinical and ultrasound follow-up of five children with
HS who had cholecystotomy with stone removal at the time of
splenectomy. In four of the cases there were no stones, but in
one an asymptomatic stone was discovered. However, the
follow-up period was short (1–2 years), and there have been
no further reports of systematic follow-up of this cohort. In a
study of 17 patients, Sandler et al (1999) identified that none
ª 2004 Blackwell Publishing Ltd, 126, 455–474
of their cohort of patients had shown either overt or subclinical gall bladder disease after a median follow-up of
18 years.
The surgical management of HS is thus largely founded on
‘custom and practice’ (level III and IV evidence) rather than
the evidence of controlled trials. There are a number of
questions that could be answered by well-designed observational studies, although to have any statistical meaning they
would have to be multi-centred.
Recommendations
1 There is no evidence that one surgical approach is
superior to another, and the choice of the laparoscopic
route would be dependent on the availability of
appropriately trained surgeons, and suitable equipment
(level III evidence, grade B recommendation).
2 Partial splenectomy may be of benefit in very young
children with severe disease, but it is likely that further
surgery may need to be undertaken for either recurrence
of haematological problems or symptomatic cholelithiasis (level IV evidence, grade C recommendation).
3 In children undergoing splenectomy for symptoms of
cholelithiasis, the gall bladder should be removed. If
stones are an incidental finding, a cholecystostomy may
be sufficient, although follow-up to include ultrasound
examination of the gall bladder will be necessary (level
III evidence, grade B recommendation).
4 In children who require cholecystectomy for symptoms
of gallstones, the spleen should always be removed as
well, otherwise the risk of stones persists (level IV
evidence, grade C recommendation).
5 When splenectomy is indicated, ideally it should be
performed after the age of 6 years, but before puberty
(level IV evidence, grade C recommendation).
What is the risk of late post splenectomy thrombosis?
Adults undergoing splenectomy should receive standard
thromboprophylaxis where indicated. Splenectomy is usually
followed by a reactive thrombocytosis that may be quite
spectacular in children, with counts rising to more than
1000 · 109/l (Hirsh & Dacie, 1966; Boxer et al, 1978; Coon
et al, 1978). Current evidence suggests that the only individuals in whom there is an increased risk of late (i.e. not
related to the surgery itself) thrombosis after splenectomy are
those with myeloproliferative disorders (Gordon et al, 1978),
or persisting anaemia with abnormal red blood cells (Hirsh &
Dacie, 1966), and has been demonstrated for beta thalassaemia intermedia (Cappellini et al, 2000) and for forms of
hereditary stomatocytosis (Delaunay et al, 1999). It is therefore vital to differentiate these membrane disorders from HS
and to avoid splenectomy. Thrombotic events have anecdo467
Guideline
tally been reported in patients with HS (Nikol et al, 1997;
Hayag-Barin et al, 1998), but the frequency of inherited
thrombotic risk factors is high enough (particularly factor V
Leiden) in northern Europeans to suggest that these will
occur by chance in association with HS. There is a suggestion
that splenectomy may be associated with the subsequent
development of pulmonary hypertension (Hoeper et al,
1999). There is some recent evidence from animal models
that there may be a risk of thrombosis related to the
abnormal red cells in mice, with both HS and HE but this
has not, as yet, been demonstrated in humans. Moreover, the
thrombosis risk was abolished in these mice by the infusion
of a small number of normal stem cells (Wandersee et al,
2001).
Recommendations
1 There is no indication for extended thrombosis prophylaxis after splenectomy in patients with HS. Adults
should receive perioperative thromboprophylaxis in the
usual way.
2 Splenectomy should be avoided in patients with some
forms of hereditary stomatocytosis (level III evidence,
grade B recommendations).
Other complications of HS and how they should
be managed
Aplastic crisis
Hereditary spherocytosis that has previously been silent and
undiagnosed may present typically in childhood with severe
anaemia caused by parvovirus B19 infection (level III
evidence). A sick febrile child is admitted with severe pallor,
typically with a maculopapular rash on the cheeks (‘slapped
cheeks’), and often with diarrhoea and vomiting. Spherocytes on the blood film together with a firm moderately
enlarged spleen are important clues; a mild accompanying
thrombocytopenia and leucopenia are common and resolve
with the infection. Anaemia may be severe, requiring blood
transfusion. The aplastic crisis typically lasts 10–14 d.
Parvovirus was recognized as the cause of aplastic crises in
sickle cell disease in 1981–82, and is now known to be
responsible for such episodes in any form of inherited
haemolysis, including HS (Kelleher et al, 1983). It is helpful
to confirm the diagnosis by documenting raised parvovirus
IgM titres in the acute illness, because the parents can then
be reassured that this is unlikely to happen again. Although
other viral illnesses cause some decompensation in HS, the
anaemia is very rarely as profound as that seen in acute
parvovirus infections. Other family members with HS should
be warned of the risk of catching this contagious viral
infection.
468
Recommendations
Where clinical suspicion exists, it may be useful to
serologically document infection with parvovirus B19.
Some children with HS and parvovirus infection may
require blood transfusion, but this should be kept to a
minimum and be guided by clinical features, as complete
recovery usually occurs within 10–14 d.
Rare complications
Growth failure may occur in severe cases, with marrow
expansion and skeletal deformities. It is good practice to
monitor height and weight serially. Leg ulcers have been
described. Several unsplenectomized adult patients have been
reported with extramedullary haemopoiesis, occasionally as the
presenting feature of the disease. (Mulder et al, 1975; Petit &
Estany, 1987; Bastion et al, 1990; Pulsoni et al, 1992; Xiros
et al, 2001; Granjo et al, 2002; Sutton et al, 2003).
Conclusions
The methodology of diagnosis and clinical management of HS
has changed over the past three decades. For many years, the
diagnosis was usually followed almost automatically by splenectomy. The diagnosis was often all too evident, but difficult
cases were rarely made easier by performing an osmotic
fragility test. There then followed a period when splenectomy
fell from favour because of the risks of postsplenectomy sepsis.
However, presplenectomy immunization against S. pneumoniae
(together with Hib and meningitis C vaccines if not already
given as part of routine vaccination) together with prophylactic antibiotic therapy has reduced, although not totally
eradicated, the risk. Postsplenectomy, there is no additional
risk of thrombosis unless the patient has some other significant
red cell disorder. Where indicated, thromboprophylaxis should
be undertaken as for other abdominal surgery in adults.
The indications for splenectomy have become somewhat
clearer and it is now possible to classify the majority of cases of
HS into mild, moderate and severe. Splenectomy will be of
benefit in all cases with severe and some cases with moderate
HS, but is not usually necessary in mild cases. However, the
final decision will rest on consultation between the family and
clinician.
Research priorities
1 Further evaluation of the EMA binding test in the district
general hospital setting is required.
2 The necessity for folate therapy needs to be reassessed in
cases with mild haemolysis. Folate levels could be systematically measured with the full blood count at routine visits
together with a record of folate therapy or absence of it.
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
3 The current Department of Health advice concerning
postsplenectomy sepsis needs review in the light of emerging pneumococcal penicillin resistance and poor compliance with long-term penicillin therapy.
4 Further research is required to establish which pneumococcal vaccine should be used, and how often it needs
repeating, and whether measurement of pneumococcal
antibodies has any role.
5 Collaborative studies of the surgical management of HS that
compare standard versus laparoscopic splenectomy would
be helpful. In addition, further information is required from
long-term studies of asymptomatic pigment gallstones – do
they inevitably result in symptoms, or can patients be left
alone, particularly in those in whom splenectomy would not
otherwise be indicated? Is it safe to do a cholecystostomy
alone? (Is there a role for national societies to collaborate
here?) From a patient perspective, the research priorities
centre on surgical intervention.
6 To determine whether or not abnormal cation efflux,
detected in red cells of some HS patients, is a consistent
indicator of increasing risk of thromboembolism postsplenectomy. As abnormal cation efflux has been reported in some HS patients, could there be a subset of
patients who present overlapping features between HS
and the red cell disorders associated with hereditary
stomatocytosis?
7 An audit should be performed on all splenectomized
patients to determine whether there is a higher incidence
of thromboembolism in HS patients than in other groups of
splenectomized individuals. It would be ideal to have a
national splenectomy register to collect this and other
information regarding postsplenectomy sepsis.
Disclaimer
While the advice and information in these guidelines is
believed to be true and accurate at the time of going to press,
neither the authors, the British Society for Haematology nor
the publishers accept any legal responsibility for the content of
these guidelines.
Source of funding
None.
Conflict of interest
None of the authors have received any funding from
companies related to the disorder or its treatment.
References
Agre, P., Asimos, A., Casella, J.F. & McMillan, C. (1986) Inheritance
pattern and clinical response to splenectomy as a reflection of ery-
ª 2004 Blackwell Publishing Ltd, 126, 455–474
throcyte spectrin deficiency in hereditary spherocytosis. New England Journal of Medicine, 315, 1579–1583.
Alloisio, N., Maillet, P., Carre, G., Texier, P., Vallier, A., Baklouti,
F., Philippe, N. & Delaunay, J. (1996) Hereditary spherocytosis
with band 3 deficiency. Association with a nonsense mutation of
the band 3 gene (allele Lyon), and aggravation by a low-expression allele occurring in trans (allele Genas). Blood, 88, 1062–
1069.
Alloisio, N., Texier, P., Vallier, A., Ribeiro, M.L., Morle, L., Bozon, M.,
Bursaux, E., Maillet, P., Goncalves, P., Tanner, M.J., Tamagnini, G.
& Delaunay, J. (1997) Modulation of clinical expression and band 3
deficiency in hereditary spherocytosis. Blood, 90, 414–420.
Bader-Meunier, B., Gauthier, F., Archambaud, F., Cynober, T., Mielot,
F., Dommergues, J.P., Warszawski, J., Mohandas, N. & Tchernia, G.
(2001) Long-term evaluation of the beneficial effect of subtotal
splenectomy for management of hereditary spherocytosis. Blood, 97,
399–403.
Bagdasarian, R.W., Bolton, J.S., Bowen, J.C., Fuhrman, G.M. &
Richardson, W.S. (2000) Steep learning curve of laparoscopic splenectomy. Journal of Laparoendoscopic and Advanced Surgical Techniques, 10, 319–323.
Baird, R.N., Macpherson, A.I.S. & Richmond, J. (1971) Red-blood-cell
survival after splenectomy in congenital spherocytosis. Lancet, ii,
1060–1061.
Bardill, B., Mengis, C., Tschopp, M. & Wuillemin, W.A. (2003) Severe
IgA-mediated auto-immune haemolytic anaemia in a 48-yr-old
woman. European Journal of Haematology, 70, 60–63.
Bastion, Y., Coiffier, B., Felman, P., Assouline, D., Tigaud, J.D.,
Espinouse, D. & Bryon, P.A. (1990) Massive mediastinal extramedullary hematopoiesis in hereditary spherocytosis: a case report.
American Journal of Hematology, 35, 263–265.
BCSH (1996) Guidelines for the prevention and treatment of infection
in patients with an absent or dysfunctional spleen: Working Party of
the British Committee for Standards in Haematology Clinical
Haematology Task Force. British Medical Journal, 312, 430–434.
Boivin, P., Galand, C., Devaux, I., Lecomte, M.C., Garbarz, M. &
Dhermy, D. (1993) Spectrin alpha IIa variant in dominant and nondominant spherocytosis. Human Genetics, 92, 153–156.
Bolton-Maggs, P.H.B. (2004) Hereditary spherocytosis – new guidelines. Archives of Disease in Childhood, in press.
Boxer, M.A., Braun, J. & Ellman, L. (1978) Thromboembolic risk of
postsplenectomy thrombocytosis. Archives of Surgery, 113, 808–809.
Bracher, N.A., Lyons, C.A., Wessels, G., Mansvelt, E. & Coetzer, T.L.
(2001) Band 3 Cape Town (E90K) causes severe hereditary spherocytosis in combination with band 3 Prague III. British Journal of
Haematology, 113, 689–693.
Bucx, M.J., Breed, W.P. & Hoffmann, J.J. (1988) Comparison of
acidified glycerol lysis test, Pink test and osmotic fragility test in
hereditary spherocytosis: effect of incubation. European Journal of
Haematology, 40, 227–231.
Cappellini, M.D., Robbiolo, L., Bottasso, B.M., Coppola, R., Fiorelli, G.
& Mannucci, A.P. (2000) Venous thromboembolism and
hypercoagulability in splenectomized patients with thalassaemia
intermedia. British Journal of Haematology, 111, 467–473.
Caprotti, R., Franciosi, C., Romano, F., Codecasa, G., Musco, F.,
Motta, M. & Uggeri, F. (1999) Combined laparoscopic splenectomy
and cholecystectomy for the treatment of hereditary spherocytosis: is
it safe and effective? Surgical Laparoscopy Endoscopy and Percutaneous Techniques, 9, 203–206.
469
Guideline
Chapman, R.G. & McDonald, L.L. (1968) Red cell life span after
splenectomy in hereditary spherocytosis. Journal of Clinical
Investigation, 47, 2263–2267.
Chilcote, R.R., Le Beau, M.M., Dampier, C., Pergament, E., Verlinsky,
Y., Mohandas, N., Frischer, H. & Rowley, J.D. (1987) Association of
red cell spherocytosis with deletion of the short arm of chromosome
8. Blood, 69, 156–159.
Clark, M.R., Mohandas, N.M. & Shohet, S.B. (1983) Osmotic gradient
ektacytometry: comprehensive characterization of the red cell
volume and surface maintenance. Blood, 61, 899–910.
Cohen, H., Walker, H., Delhanty, J.D., Lucas, S.B. & Huehns, E.R.
(1991) Congenital spherocytosis, B19 parvovirus infection and
inherited interstitial deletion of the short arm of chromosome 8.
British Journal of Haematology, 78, 251–257.
Coon, W.W., Penner, J., Clagett, P. & Eos, N. (1978) Deep venous
thrombosis and postsplenectomy thrombocytosis. Archives of
Surgery, 113, 429–431.
Costa, F.F., Agre, P., Watkins, P.C., Winkelmann, J.C., Tang, T.K., John,
K.M., Lux, S.E. & Forget, B.G. (1990) Linkage of dominant hereditary
spherocytosis to the gene for the erythrocyte membrane-skeleton
protein ankyrin. New England Journal of Medicine, 323, 1046–1050.
Csendes, A., Smok, G., Burdiles, P., Diaz, J.-C., Maluenda, F. &
Korn, O. (1998) Histological findings of gallbladder mucosa in 95
control subjects and 80 patients with asymptomatic gallstones. Digestive Diseases and Sciences, 43, 931–934.
Curran, T.J., Foley, M.I., Swanstrom, L.L. & Campbell, T.J. (1998)
Laparoscopy improves outcomes for pediatric splenectomy. Journal
of Pediatric Surgery, 33, 1498–1500.
Cynober, T., Mohandas, N. & Tchernia, G. (1996) Red cell abnormalities in hereditary spherocytosis: relevance to diagnosis and
understanding of the variable expression of clinical severity. Journal
of Laboratory and Clinical Medicine, 128, 259–269.
Dacie, J.V., Lewis, S.M. & Luzatto, L. (1991) Investigation of the
hereditary haemolytic anaemias: membrane and enzyme abnormalities. In: Practical Haematology (ed. by J.V. Dacie & S.M. Lewis),
pp. 195–225. Churchill Livingstone, Edinburgh.
Dallman, P.R. & Siimes, M.A. (1979) Percentile curves for hemoglobin
and red cell volume in infancy and childhood. Journal of Pediatrics,
94, 26–31.
Danielson, P.D., Shaul, D.B., Phillips, J.D., Stein, J.E. & Anderson, K.D.
(2000) Technical advances in pediatric laparoscopy have had a
beneficial impact on splenectomy. Journal of Pediatric Surgery, 35,
1578–1581.
Davies, J.M., Barnes, R. & Milligan, D. (2002) Update of guidelines for
the prevention and treatment of infection in patients with an absent
or dysfunctional spleen. Clinical Medicine: Journal of the Royal
College of Physicians of London, 2, 440–443.
De Franceschi, L., Olivieri, O., Miraglia del Giudice, E., Perrotta, S.,
Sabato, V., Corrocher, R. & Iolascon, A. (1997) Membrane cation
and anion transport activities in erythrocytes of hereditary spherocytosis: effects of different membrane protein defects. American
Journal of Hematology, 55, 121–128.
Delaunay, J., Alloisio, N., Morle, L., Baklouti, F., Dalla Venezia, N.,
Maillet, P. & Wilmotte, R. (1996) Molecular genetics of hereditary
elliptocytosis and hereditary spherocytosis. Annals of Genetics, 39,
209–221.
Delaunay, J., Stewart, G. & Iolascon, A. (1999) Hereditary dehydrated
and overhydrated stomatocytosis: recent advances. Current Opinion
in Hematology, 6, 110–114.
470
Delhommeau, F., Cynober, T., Schischmanoff, P.O., Rohrlich, P.,
Delaunay, J., Mohandas, N. & Tchernia, G. (2000) Natural history of
hereditary spherocytosis during the first year of life. Blood, 95, 393–
397.
Deodhar, H.A., Marshall, R.J. & Barnes, J.N. (1993) Increased risk of
sepsis after splenectomy. British Medical Journal, 307, 1408–1409.
Dhermy, D., Galand, C., Bournier, O., Boulanger, L., Cynober, T.,
Schismanoff, P.O., Bursaux, E., Tchernia, G., Boivin, P. & Garbarz,
M. (1997) Heterogenous band 3 deficiency in hereditary spherocytosis related to different band 3 gene defects [published erratum
appears in British Journal of Haematology 1997, 99, 474]. British
Journal of Haematology, 98, 32–40.
Dhermy, D., Steen-Johnsen, J., Bournier, O., Hetet, G., Cynober, T.,
Tchernia, G. & Grandchamp, B. (2000) Coinheritance of two alphaspectrin gene defects in a recessive spherocytosis family. Clinical and
Laboratory Haematology, 22, 329–336.
Eber, S.W., Armbrust, R. & Schroter, W. (1990) Variable clinical
severity of hereditary spherocytosis: relation to erythrocytic spectrin
concentration, osmotic fragility and autohemolysis. Journal of
Pediatrics, 177, 409–411.
Eber, S.W., Pekrun, A., Neufeldt, A. & Schroter, W. (1992) Prevalence
of increased osmotic fragility of erythrocytes in German blood
donors: screening using a modified glycerol lysis test. Annals of
Hematology, 64, 88–92.
Eber, S.W., Gonzalez, J.M., Lux, M.L., Scarpa, A.L., Tse, W.T., Dornwell, M., Herbers, J., Kugler, W., Ozcan, R., Pekrun, A., Gallagher,
P.G., Schroter, W., Forget, B.G. & Lux, S.E. (1996) Ankyrin-1
mutations are a major cause of dominant and recessive hereditary
spherocytosis. Nature Genetics, 13, 214–218.
Eber, S.W., Langendorfer, C.M., Ditzig, M., Reinhardt, D., Stohr, G.,
Soldan, W., Schroter, W. & Tchernia, G. (1999) Frequency of very
late fatal sepsis after splenectomy for hereditary spherocytosis:
impact of insufficient antibody response to pneumococcal infection.
Annals of Hematology, 78, 524–528.
Ein, S.H., Shandling, B., Simpson, J.S., Stephens, C.A., Bandi, S.K.,
Biggar, W.D. & Freedman, M.H. (1977) The morbidity and mortality of splenectomy in childhood. Annals of Surgery, 185, 307–310.
Eraklis, A.J. & Filler, R.M. (1972) Splenectomy in childhood: a review
of 1413 cases. Journal of Pediatric Surgery, 7, 382–388.
Esposito, C., Corcione, F., Ascione, G., Garipoli, V., Di Pietto, F. & De
Pasquale, M. (1998) Splenectomy in childhood. The laparoscopic
approach. Surgical Endoscopy, 12, 1445–1448.
Evans, D.I. (1985) Postsplenectomy sepsis 10 years or more after
operation. Journal of Clinical Pathology, 38, 309–311.
Farah, R.A., Rogers, Z.R., Thompson, W.R., Hicks, B.A., Guzzetta, P.C.
& Buchanan, G.R. (1997) Comparison of laparoscopic and open
splenectomy in children with hematologic disorders. Journal of
Pediatrics, 131, 41–46.
Friedman, E.W., Williams, J.C. & Van Hook, L. (1988) Hereditary
spherocytosis in the elderly. American Journal of Medicine, 84, 513–
516.
Gallagher, P.G., Weed, S.A., Tse, W.T., Benoit, L., Morrow, J.S.,
Marchesi, S.L., Mohandas, N. & Forget, B.G. (1995) Recurrent fatal
hydrops fetalis associated with a nucleotide substitution in the
erythrocyte beta-spectrin gene. Journal of Clinical Investigation, 95,
1174–1182.
Gallagher, P., Forget, B. & Lux, S. (1998) Disorders of the erythrocyte
membrane. In: Nathan and Oski’s Hematology of Infancy and
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
Childhood (ed. by D. Nathan & S. Orkin), Vol. 1, pp. 544–664.
Saunders, Philadelphia.
Gallagher, P.G., Ferreira, J.D., Costa, F.F., Saad, S.T. & Forget, B.G.
(2000) A recurrent frameshift mutation of the ankyrin gene
associated with severe hereditary spherocytosis. British Journal of
Haematology, 111, 1190–1193.
Garbarz, M., Bibas, D., Cynober, T., Galand, C., Bournier, O., Devaux, I.,
Tchernia, G. & Dhermy, D. (1996) Search for the candidate genes in
dominant hereditary spherocytosis using linkage analysis. Comptes
Rendus de l’ Academie des Sciences, Paris, Sciences de la Vie III, 319,
913–919.
Garbarz, M., Galand, C., Bibas, D., Bournier, O., Devaux, I.,
Harousseau, J.L., Grandchamp, B. & Dhermy, D. (1998) A 5¢ splice
region G–>C mutation in exon 3 of the human beta-spectrin gene
leads to decreased levels of beta-spectrin mRNA and is responsible
for dominant hereditary spherocytosis (spectrin Guemene-Penfao).
British Journal of Haematology, 100, 90–98.
Gehlbach, S.H. & Cooper, B.A. (1970) Haemolytic anaemia in
infectious mononucleosis due to inapparent congenital spherocytosis. Scandinavian Journal of Haematology, 7, 141–144.
Godal, H.C. & Heisto, H. (1981) High prevalence of increased osmotic
fragility of red blood cells among Norwegian blood donors.
Scandinavian Journal of Haematology, 27, 30–34.
Goodman, S.R., Shiffer, K.A., Casoria, L.A. & Eyster, M.E. (1982)
Identification of the molecular defect in the erythrocyte membrane
skeleton of some kindreds with hereditary spherocytosis. Blood, 60,
772–784.
Gordon, D.H., Schaffner, D., Bennett, J.M. & Schwartz, S.I. (1978)
Postsplenectomy thrombocytosis: its association with mesenteric,
portal, and/or renal vein thrombosis in patients with myeloproliferative disorders. Archives of Surgery, 113, 713–715.
Gotz, M., Granditsch, G., Helmer, F., Howanietz, L. & Wagner, I.U.
(1977) Cholelithiasis in children and adolescents (authors transl)
Wien Klin Wochenschr, 89, 686–690.
Gracie, W.A. & Ransohoff, D.F. (1982) The natural history of silent
gallstones: the innocent gallstone is not a myth. New England Journal
of Medicine, 307, 798–800.
Granjo, E., Bauerle, R., Sampaio, R., Manata, P., Torres, N. & Quintanilha, A. (2002) Extramedullary hematopoiesis in hereditary
spherocytosis deficient in ankyrin: a case report. International
Journal of Hematology, 76, 153–156.
Haines, P.G., Jarvis, H.G., King, S., Noormohamed, F.H., Chetty,
M.C., Fisher, J., Hill, P., Nicolaou, A. & Stewart, G.W. (2001)
Two further British families with the ‘cryohydrocytosis’ form of
hereditary stomatocytosis. British Journal of Haematology, 113,
932–937.
Hassoun, H., Vassiliadis, J.N., Murray, J., Njolstad, P.R., Rogus, J.J.,
Ballas, S.K., Schaffer, F., Jarolim, P., Brabec, V. & Palek, J. (1997)
Characterization of the underlying molecular defect in hereditary
spherocytosis associated with spectrin deficiency. Blood, 90, 398–
406.
Hayag-Barin, J.E., Smith, R.E. & Tucker, F.C., Jr (1998) Hereditary
spherocytosis, thrombocytosis, and chronic pulmonary emboli: a
case report and review of the literature. American Journal of
Hematology, 57, 82–84.
Hirsh, J. & Dacie, J.V. (1966) Persistent post-splenectomy thrombocytosis and thrombo-embolism: a consequence of continuing
anaemia. British Journal of Haematology, 12, 44–53.
ª 2004 Blackwell Publishing Ltd, 126, 455–474
Hoeper, M.M., Niedermeyer, J., Hoffmeyer, F., Flemming, P. & Fabel,
H. (1999) Pulmonary hypertension after splenectomy? Annals of
Internal Medicine, 130, 506–509.
Holcomb, G.W., III (1998) Laparoscopic cholecystectomy. Seminars of
Laparoscopic Surgery, 5, 2–8.
Holdsworth, R.J., Irving, A.D. & Cuschieri, A. (1991) Postsplenectomy
sepsis and its mortality rate: actual versus perceived risks. British
Journal of Surgery, 78, 1031–1038.
Iolascon, A., Miraglia del Giudice, E., Perrotta, S., Alloisio, N., Morle,
L. & Delaunay, J. (1998) Hereditary spherocytosis: from clinical to
molecular defects. Haematologia, 83, 240–257.
Jarolim, P., Rubin, H.L., Brabec, V. & Palek, J. (1995) Comparison of
the ankyrin (AC)n microsatellites in genomic DNA and mRNA
reveals absence of one ankyrin mRNA allele in 20% of patients with
hereditary spherocytosis. Blood, 85, 3278–3282.
Jarolim, P., Murray, J.L., Rubin, H.L., Taylor, W.M., Prchal, J.T.,
Ballas, S.K., Snyder, L.M., Chrobak, L., Melrose, W.D., Brabec, V. &
Palek, J. (1996) Characterization of 13 novel band 3 gene defects in
hereditary spherocytosis with band 3 deficiency. Blood, 88, 4366–
4374.
Jensson, O., Jonasson, J.L. & Magnusson, S. (1977) Studies on
hereditary spherocytosis in Iceland. Acta Medica Scandinavica, 201,
187–195.
Jugenburg, M., Haddock, G., Freedman, M.H., Ford-Jones, L. & Ein,
S.H. (1999) The morbidity and mortality of pediatric splenectomy:
does prophylaxis make a difference? Journal of Pediatric Surgery, 34,
1064–1067.
Keenan, R.D., Boswell, T. & Milligan, D.W. (1999) Do postsplenectomy patients take their penicillin. British Journal of
Haematology, 105, 509–510.
Kelleher, J.F., Luban, N.L., Mortimer, P.P. & Kamimura, T. (1983)
Human serum ‘parvovirus’: a specific cause of aplastic crisis in
children with hereditary spherocytosis. Journal of Pediatrics, 102,
720–722.
Kerr, R., Rawlinson, T.S. & Cachia, P.G. (2000) Direct antiglobulin test
negative, non-spherocytic autoimmune haemolytic anaemia. Clinical and Laboratory Haematology, 22, 365–367.
King, M.J., Behrens, J., Rogers, C., Flynn, C., Greenwood, D. &
Chambers, K. (2000) Rapid flow cytometric test for the diagnosis of
membrane cytoskeleton-associated haemolytic anaemia. British
Journal of Haematology, 111, 924–933.
King, M.-J., Chapman, L., Mackinnon, H., Mills, W., PsiachouLeonnard, E. & Murrin, R. (2003) Examination of flow cytometric
histograms of eosin-5-maleimide labelled red cells can assist in
differential diagnosis of membranopathy. British Journal of Haematology, 121(Suppl. 1), 75.
King, M.-J., Smythe, J. & Mushens, R. (2004) Eosin-5-maleimide
binding to band 3 and Rh-related proteins forms the basis of a
screening test for hereditary spherocytosis. British Journal of Haematology, 124, 106–113.
Kohler, H., Meynell, M. & Cooke, W. (1960) Spherocytic anaemia
complicated by megaloblastic anaemia of pregnancy. British Medical
Journal, 1, 779–781.
Konradsen, H.B. & Henrichsen, J. (1991) Pneumococcal infections in
splenectomized children are preventable. Acta Paediatrica Scandinavica, 80, 423–427.
Korones, D. & Pearson, H.A. (1989) Normal erythrocyte osmotic fragility in hereditary spherocytosis. Journal of Pediatrics, 114, 264–266.
471
Guideline
Lefrere, J.J., Courouce, A.M., Girot, R., Bertrand, Y. & Soulier, J.P.
(1986) Six cases of hereditary spherocytosis revealed by human
parvovirus infection. British Journal of Haematology, 62, 653–658.
Lindenbaum, J. & Klipstein, F.A. (1963) Folic acid deficiency in sicklecell anemia. New England Journal of Medicine, 269, 875–882.
Mackenzie, F., Elliot, D., Eastcott, H., Hughes-Jones, N., Barkhan, P. &
Mollinson, P. (1962) Relapse in hereditary spherocytosis with proven splenunculus. Lancet, 1, 1102–1104.
Marchetti, M., Quaglini, S. & Barosi, G. (1998) Prophylactic splenectomy and cholecystectomy in mild hereditary spherocytosis:
analyzing the decision in different clinical scenarios. Journal of
Internal Medicine, 244, 217–226.
McLane, N.J., Grizzard, W.S., Kousseff, B.G., Hartmann, R.C. & Sever,
R.J. (1984) Angioid streaks associated with hereditary spherocytosis.
American Journal of Ophthalmology, 97, 444–449.
Michaels, L.A., Cohen, A.R., Zhao, H., Raphael, R.I. & Manno, C.S.
(1997) Screening for hereditary spherocytosis by use of automated
erythrocyte indexes. Journal of Pediatrics, 130, 957–960.
Miraglia del Giudice, E., Iolascon, A., Pinto, L., Nobili, B. & Perrotta, S.
(1994) Erythrocyte membrane protein alterations underlying clinical
heterogeneity in hereditary spherocytosis. British Journal of Haematology, 88, 52–55.
Miraglia del Giudice, E., Lombardi, C., Francese, M., Nobili, B.,
Conte, M.L., Amendola, G., Cutillo, S., Iolascon, A. & Perrotta, S.
(1998) Frequent de novo monoallelic expression of beta-spectrin
gene (SPTB) in children with hereditary spherocytosis and isolated spectrin deficiency. British Journal of Haematology, 101,
251–254.
Miraglia del Giudice, E., Perrotta, S., Nobili, B., Specchia, C., d’Urzo,
G. & Iolascon, A. (1999) Coinheritance of Gilbert syndrome
increases the risk for developing gallstones in patients with hereditary spherocytosis. Blood, 94, 2259–2262.
Miraglia del Giudice, E., Nobili, B., Francese, M., D’Urso, L., Iolascon,
A., Eber, S. & Perrotta, S. (2001) Clinical and molecular evaluation
of non-dominant hereditary spherocytosis. British Journal of Haematology, 112, 42–47.
Mohandas, N., Clark, M.R., Jacobs, M.S. & Shohet, S.B. (1980) Analysis of factors regulating erythrocyte deformability. Journal of
Clinical Investigation, 66, 563–573.
Mohandas, N., Kim, Y.R., Tycko, D.H., Orlik, J., Wyatt, J. & Groner, W.
(1986) Accurate and independent measurement of volume and hemoglobin concentration of individual red cells by laser light scattering. Blood, 68, 506–513.
Morton, N., MacKinney, A., Kosowe, N., Schilling, R. & Gray, M.
(1962) Genetics of spherocytosis. American Journal of Human Genetics, 14, 170–184.
Mulder, H., Schlangen, J.T. & van Voorthuisen, A.E. (1975) Extramedullary hematopoiesis in the posterior mediastinum. Radiology
Clinics (Basel), 44, 550–556.
Nikol, S., Huehns, T.Y., Kiefmann, R. & Hofling, B. (1997) Excessive
arterial thrombus in spherocytosis. A case report. Angiology, 48,
743–748.
Pajor, A., Lehoczky, D. & Szakacs, Z. (1993) Pregnancy and hereditary
spherocytosis. Report of 8 patients and a review. Archives of Gynecology and Obstetrics, 253, 37–42.
Palek, J. & Sahr, K.E. (1992) Mutations of the red blood cell membrane
proteins: from clinical evaluation to detection of the underlying
genetic defect. Blood, 80, 308–330.
472
Parpart, A.K., Lorenz, P.B., Parpart, E.R., Gregg, J.R. & Chase, A.M.
(1947) The osmotic resistance (fragility) of human red cells. Journal
of Clinical Investigation, 26, 636–640.
Pati, A.R., Patton, W.N. & Harris, R.I. (1989) The use of the technicon
H1 in the diagnosis of hereditary spherocytosis. Clinical and
Laboratory Haematology, 11, 27–30.
Pedersen, F.K. (1983) Postsplenectomy infections in Danish children
splenectomized 1969–1978. Acta Paediatrica Scandinavica, 72, 589–
595.
Petit, J.J. & Estany, C. (1987) Mediastinal extramedullary erythropoiesis in hereditary spherocytosis. Clinical and Laboratory Haematology, 9, 327–332.
Pilar Ricard, M. & Gilanz, F. (1996) Assessment of the severity of
hereditary spherocytosis using routine haematological data obtained
with dual angle laser scattering cytometry. Clinical and Laboratory
Haematology, 18, 75–78.
Pinto, L., Iolascon, A., Miraglia del Giudice, E., Matarese, S.M.R.,
Nobili, B. & Perrotta, S. (1995) The Italian survey on hereditary
spherocytosis. International Journal of Pediatric Hematology and
Oncology, 2, 43–47.
Pulsoni, A., Ferrazza, G., Malagnino, F., Maurillo, L., Pescarmona, E.,
Picardi, A., Rendina, E.A. & Amadori, S. (1992) Mediastinal extramedullary hematopoiesis as first manifestation of hereditary spherocytosis. Annals of Hematology, 65, 196–198.
Reid, M.M. (1994) Splenectomy, sepsis, immunisation, and guidelines.
Lancet, 344, 970–971.
Reinhart, W.H., Wyss, E.J., Arnold, D. & Ott, P. (1994) Hereditary
spherocytosis associated with protein band 3 defect in a Swiss kindred. British Journal of Haematology, 86, 147–155.
Reliene, R., Mariani, M., Zanella, A., Reinhart, W.H., Ribeiro, M.L., del
Giudice, E.M., Perrotta, S., Iolascon, A., Eber, S. & Lutz, H.U. (2002)
Splenectomy prolongs in vivo survival of erythrocytes differently in
spectrin/ankyrin- and band 3-deficient hereditary spherocytosis.
Blood, 100, 2208–2215.
Rescorla, F.J. & Grosfeld, J.L. (1992) Cholecystitis and cholelithiasis in
children. Seminars of Pediatric Surgery, 1, 98–106.
Rescorla, F.J., Engum, S.A., West, K.W., Tres Scherer, L.R., III, Rouse,
T.M. & Grosfeld, J.L. (2002) Laparoscopic splenectomy has become
the gold standard in children. American Surgery, 68, 297–301; discussion 301–302.
Rice, H.E., Oldham, K.T., Hillery, C.A., Skinner, M.A., O’Hara, S.M. &
Ware, R.E. (2003) Clinical and hematologic benefits of partial
splenectomy for congenital hemolytic anemias in children. Annals of
Surgery, 237, 281–288.
Robertson, J.F., Carachi, R., Sweet, E.M. & Raine, P.A. (1988) Cholelithiasis in childhood: a follow-up study. Journal of Pediatric Surgery,
23, 246–249.
Rosenblatt, D.S. & Hoffbrand, A.V. (1999) Megaloblastic anemia and
disorders of cobalamin and folate metabolism. In: Pediatric Hematology (ed. by J.S. Lilleyman, I.M. Hann & V.S. Blanchette), pp. 167–
184. Churchill Livingstone, London.
Rutkow, I.M. (1981) Twenty years of splenectomy for hereditary
spherocytosis. Archives of Surgery, 116, 306–308.
Rybicki, A.C., Schwartz, R.S., Hustedt, E.J. & Cobb, C.E. (1996) Increased rotational mobility and extractability of band 3 from protein
4.2-deficient erythrocyte membranes: evidence of a role for protein
4.2 in strengthening the band 3-cytoskeleton linkage. Blood, 88,
2745–2753.
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474
Guideline
Sandler, A., Winkel, G., Kimura, K. & Soper, R. (1999) The role of
prophylactic cholecystectomy during splenectomy in children with
hereditary spherocytosis. Journal of Pediatric Surgery, 34, 1077–
1078.
Schroter, W. & Kahsnitz, E. (1983) Diagnosis of hereditary spherocytosis in newborn infants. Journal of Pediatrics, 103, 460–463.
Smith, B.M., Schropp, K.P., Lobe, T.E., Rogers, D.A., Presbury, G.J.,
Wilimas, J.A. & Wong, W.C. (1994) Laparoscopic splenectomy in
childhood. Journal of Pediatric Surgery, 29, 975–977.
Spickett, G.P., Bullimore, J., Wallis, J., Smith, S. & Saunders, P. (1999)
Northern Region asplenia register – analysis of first two years.
Journal of Clinical Pathology, 52, 424–429.
Stenson, P.D., Ball, E.V., Mort, M., Phillips, A.D., Shiel, J.A., Thomas,
N.S., Abeysinghe, S., Krawczak, M. & Cooper, D.N. (2003) Human
Gene Mutation Database (HGMD): 2003 update. Human Mutation,
21, 577–581.
Stewart, G.W. & Turner, E.J. (1999) The hereditary stomatocytoses and
allied disorders: congenital disorders of erythrocyte membrane
permeability to Na and K. Baillieres Best Practice and Research
Clinical Haematology, 12, 707–727.
Stewart, G.W., Amess, J.A., Eber, S.W., Kingswood, C., Lane, P.A.,
Smith, B.D. & Mentzer, W.C. (1996) Thrombo-embolic disease after
splenectomy for hereditary stomatocytosis. British Journal of
Haematology, 93, 303–310.
Streichman, S. & Gescheidt, Y. (1998) Cryohemolysis for the detection
of hereditary spherocytosis: correlation studies with osmotic fragility
and autohemolysis. American Journal of Hematology, 58, 206–212.
Summerfield, G.P. & Wyatt, G.P. (1985) Human parvovirus infection
revealing hereditary spherocytosis. Lancet, 2, 1070.
Sutton, C.D., Garcea, G., Marshall, L.J., Lloyd, T.D., De Alwis, C. &
Lewis, M.H. (2003) Pelvic extramedullary haematopoiesis associated
with hereditary spherocytosis. European Journal of Haematology, 70,
326–329.
Tanoue, K., Okita, K., Akahoshi, T., Konishi, K., Gotoh, N., Tsutsumi,
N., Tomikawa, M. & Hashizume, M. (2002) Laparoscopic splenectomy for hematologic diseases. Surgery, 131, S318–S323.
Tchernia, G., Gauthier, F., Mielot, F., Dommergues, J.P., Yvart, J.,
Chasis, J.A. & Mohandas, N. (1993) Initial assessment of the beneficial effect of partial splenectomy in hereditary spherocytosis.
Blood, 81, 2014–2020.
Tchernia, G., Bader-Meunier, B., Berterottiere, P., Eber, S., Dommergues, J.P. & Gauthier, F. (1997) Effectiveness of partial splenectomy
in hereditary spherocytosis. Current Opinion in Hematology, 4, 136–
141.
Tchernia, G., Delhommeau, F., Perrotta, S., Cynober, T., Bader-Meunier, B., Nobili, B., Rohrlich, P., Salomon, J.L., Sagot-Bevenot, S., del
Giudice, E.M., Delaunay, J., DeMattia, D., Schischmanoff, P.O.,
Mohandas, N. & Iolascon, A. (2000) Recombinant erythropoietin
therapy as an alternative to blood transfusions in infants with
hereditary spherocytosis. Hematology Journal, 1, 146–152.
Tse, W.T. & Lux, S.E. (1999) Red blood cell membrane disorders.
British Journal of Haematology, 104, 2–13.
Waghorn, D. & Mayon-White, R. (1997) A study of 42 episodes of
overwhelming post-splenectomy infection: is guidance for asplenic
individuals being followed? Journal of Infection, 35, 289–294.
Walker, W. (1976) Splenectomy in childhood; a review in England and
Wales. British Journal of Surgery, 63, 36–43.
Wandersee, N.J., Lee, J.C., Deveau, S.A. & Barker, J.E. (2001) Reduced
incidence of thrombosis in mice with hereditary spherocytosis fol-
ª 2004 Blackwell Publishing Ltd, 126, 455–474
lowing neonatal treatment with normal hematopoietic cells. Blood,
97, 3972–3975.
Whitfield, C.F., Follweiler, J.B., Lopresti-Morrow, L. & Miller, B.A.
(1991) Deficiency of alpha-spectrin synthesis in burst-forming
units-erythroid in lethal hereditary spherocytosis. Blood, 78, 3043–
3051.
Wichterle, H., Hanspal, M., Palek, J. & Jarolim, P. (1996) Combination
of two mutant alpha spectrin alleles underlies a severe spherocytic
hemolytic anemia. Journal of Clinical Investigation, 98, 2300–2307.
Wolfe, L.C., John, K.M., Falcone, J.C., Byrne, A.M. & Lux, S.E. (1982)
A genetic defect in the binding of protein 4.1 to spectrin in a kindred
with hereditary spherocytosis. New England Journal of Medicine, 307,
1367–1374.
Xiros, N., Economopoulos, T., Papageorgiou, E., Mantzios, G. &
Raptis, S. (2001) Massive hemothorax due to intrathoracic extramedullary hematopoiesis in a patient with hereditary spherocytosis.
Annals of Hematology, 80, 38–40.
Yawata, Y., Kanzaki, A., Inoue, T., Ata, K., Wada, H., Okamoto, N.,
Higo, I., Yawata, A., Sugihara, T. & Yamada, O. (1994) Red cell
membrane disorders in the Japanese population: clinical, biochemical, electron microscopic, and genetic studies. International Journal
of Hematology, 60, 23–38.
Zanella, A., Izzo, C., Rebulla, P., Zanuso, F., Perroni, L. & Sirchia, G.
(1980) Acidified glycerol lysis test: a screening test for spherocytosis.
British Journal of Haematology, 45, 481–486.
Appendix
The definitions of the types of evidence and the grading
recommendations used in this guideline originate from the US
Agency for Health Care Policy and Research and are set out
below.
Statements of evidence
Ia Evidence obtained from the meta-analysis of randomized
controlled trials.
Ib Evidence obtained from at least one randomized controlled
trial.
IIa Evidence obtained from at least one well-designed
controlled study without randomization.
IIb Evidence obtained from at least one other type of welldesigned quasi-experimental study.
III Evidence obtained from well-designed non-experimental
descriptive studies, such as comparative studies, correlation studies and case studies.
IV Evidence obtained from expert committee reports or
opinions and/or clinical experiences of respected authorities.
Grades of recommendations
A Required at least one randomized controlled trial as part of
a body literature of overall good quality and consistency
addressing the specific recommendations (evidence levels
Ia, Ib).
473
Guideline
B Requires the available of well conducted clinical studies but
no randomized clinical trials on the topic of recommendations (evidence levels IIa, IIb, III).
474
C Requires evidence obtained from expert committee reports
or opinions and/or clinical experiences of respected authorities. Indicates an absence of directly applicable clinical
studies of good quality (evidence level IV).
ª 2004 Blackwell Publishing Ltd, British Journal of Haematology, 126, 455–474